I'.AMTU PLACE- NAMi:s IN AFRICA. 2O5 



has, oti the hitter ([uestiun, as it seems to me, so clearh' proved 

 in his work on Hottentot Sounds in Kaffir, I'ia. : that chcks 

 (with /'/; and r) are amonsrst these, or. at any rate, non-Bantu; 

 an(K tlierefore, names containinj; llieni do not directly concern 

 us in this ])aper. ( )n the other liand. to he clear, I propose 

 rapidly to trace the north and cast limit oi these ancient non- 

 Bantu place-names left hehind in the Bantu advance to south 

 and west. This is comparativel}- easy in the south-east, as far 

 north as the land occupied hy the Southern Basuto, i.e., to the 

 north marg^in of the " Conc|uerc(l Territory." The reason for 

 this ease is the presence of clicks. re])resented in Kaffir and 

 Zulu by (• q x and their \ariants. the middle or q grroup alone 

 being used in Sesut(\ Fhere is no need for me to dwell on them, 

 as they lie outside oiu- subject ; but there was need to mention 

 them, since these clicks have been so completely adopted into 

 Zulu and Kaffir that it re([uires care to note their ])resence, and 

 block out the names containing them from among the ])ure Bantu 

 names. Thus, the native name for Kingwilliamstown ( c()(>)icc) 

 begins with the regular Kaffir locative prefix c. and the name 

 of the mountain Qhoqholosing ends with ////. the normal .SiUo 

 locative ; and this is apt to withdraw attention from the fact that 

 both contain non-Bantu sounds, and should not, therefore, enter 

 into our present consideration. I hope, on another occasion, to 

 be allowed to treat more fully of these Bushman and Hottentot 

 names, but at present we must only note their existence and 

 avoid them, after determining their limits on the border of tlie 

 Bantu sphere. 



An examination of these on a large scale map reveals the 

 Xqakaka and Xora Rivers as the last names on the coast to- 

 wards the east, which contain a non-Bantu sound (excepting, of 

 course, the Portuguese and Arab names). These rivers are just 

 south of the Umtata. From thence the limit line seems to pass 

 to Ixopo ( Stuartstown ), and thence to a groujj of clicked names 

 near the Buffalo. These are N(|utu Mountain, l)el<)w the junc- 

 tion of its tributary, the Blood River, with Mcenci Hill to the 

 north, and Qudini Forest to the south-east of the Buft'alo, and 

 between Sunday's and Tugela Rivers. Oan(|ana Mountain and 

 Umqwendu. Further u]), on the east of the Drakensberg, we 

 find Incandu and lngcul)u. respectively southern and northern 

 tributaries of the Buffalo, and Umbumrana. with the tell-tale r, 

 due east of these, on latitude 27.40. 



In Basutoland, we find several on the Orange River, or 

 Sencju, and also south of it, up to its source, where Qho(|holosing 

 Mountain, already mentioned, stands as a gigantic outpost of the 

 non-Bantu names. Coming do\Nn the Orange again, we meet 

 the click not adojited into Suto, but occurring in Kaffir, in Xaba, 

 north of Herschel district, Nxaxa in Qumbu, and Mcoseli's in 

 Umzimkulu districts. Westward, throughout the whole of the 

 old colony, we find pure Hottentot and Bushman names^ dificr- 

 entlv spelt (with variations) from those adopted into Kaffir m 



